Instance segmentation is an essential part of image semantic understanding. In this paper we propose a novel cascade framework for instance segmentation. Unlike some existing methods that only output contour discrete coordinates, our approach obtains a vectorized representation of contours using periodic B-splines. In order to make better use of geometry and appearance information, we consider the global and local features of objects and introduce two types of graph structures, the star graph and circular graph, for feature extraction. Thereby, we develop a neural network, termed the mix network, to better exploit extracted features. Specifically, we first regress the spline control points to an object boundary via the mix network, then perform spline sampling to obtain the initial predictions of contours, and finally deform the predicted contours to the real contours of the objects. In addition, we add a regularization to further constrain the fairness of contour splines. Experiments show that our approach achieves 34.6% in mask mAP, Mean Average Precision, with a ResNet-101-FPN-DCN backbone on the challenging COCO benchmark, which significantly improves the performance of contour-based methods.

实例分割是图像语义理解的重要组成部分。在本文中，我们提出了一种新颖的级联框架用于实例分割。与一些仅输出轮廓离散坐标的现有方法不同，我们的方法使用周期性 B 样条获得轮廓的矢量化表示。为了更好地利用几何和外观信息，我们考虑了对象的全局和局部特征，并引入了星形图和圆形图两种图结构进行特征提取。因此，我们开发了一个神经网络，称为混合网络，以更好地利用提取的特征。具体来说，我们首先通过混合网络将样条控制点回归到对象边界，然后进行样条采样以获得轮廓的初始预测，最后将预测的轮廓变形为物体的真实轮廓。此外，我们添加了正则化以进一步限制轮廓样条的公平性。实验表明，我们的方法在具有挑战性的 COCO 基准测试中使用 ResNet-101-FPN-DCN 主干在掩码 mAP、平均平均精度上达到了 34.6%，这显着提高了基于轮廓的方法的性能。